IR spectra of as-prepared precipitates, recrystallized samples, and crystals prepared with Na-W₁₀ showed the presence of C₁₆py cations and W₁₀ anions. The single crystal X-ray structure analysis combined with the elemental analysis revealed the formula of C₁₆py-W₁₀ to be [C₅H₅N(C₁₆H₃₃)]₄[W₁₀O₃₂] (Table 1). Four C₁₆py cations (1+ charge) were associated with one W₁₀ anion (4− charge), and C₁₆py-W₁₀ did not contain any H⁺. Figure 1 shows the crystal structure of C₁₆py-W₁₀. The crystal packing consisted of alternating W₁₀ inorganic layers and C₁₆py organic layers with periodicity of 23.3 Å (Figure 1). The W₁₀ anions formed monolayers, while the hexadecyl chains of C₁₆py interdigitated to form a bilayer structure, which is a typical structure for most POM-surfactant hybrid crystals [26,27,28,29,30,31,32]. All C–C bonds in the hexadecyl chains showed anti conformation except one C–C bond (C7–C8). While both C₁₆py-W₁₀ and C₁₆pyCl·H₂O [39] contained interdigitated bilayers of C₁₆py with the pyridine ring inserted into the hydrophilic layers, the packing of hydrophilic layers was different; the W₁₀ monolayers for C₁₆py-W₁₀ and the Cl^––H₂O bilayers for C₁₆pyCl·H₂O.

The hydrophilic heads of C₁₆py penetrated into the W₁₀ inorganic monolayers and isolated each W₁₀ anion (Figure 2) in a similar way to that in the crystal of C₁₆py-hexamolybdate (C₁₆py-Mo₆) [31] or C₁₆py-α-octamolybdate (C₁₆py-α-Mo₈) [32]. On the other hand, two independent C₁₆py cations were not parallel (angle: 52.6°) without π–π stacking, different from C₁₆py-Mo₆ [31] and C₁₆py-α-Mo₈ [32]. The C–H···π(centroid) distance was 3.32 Ǻ, and the nearest C–H bond (C24–H24) was not directed to the center of the pyridine ring. The shortest interatomic distance (C3···H24, 2.88 Ǻ) between the pyridine rings was almost the same as the sum of the van der Waals radii (2.90 Ǻ). Therefore, C–H···π interaction [40,41] was hardly observed in the present C₁₆py-W₁₀.

C₁₆py-W₁₀ had two-dimensionally confined monolayers of the W₁₀ anions (Figure 2). The distance between the nearest W₁₀ anions was 3.23 Å, much shorter than that for W₁₀ crystals composed of n-butyl- [42,43] or n-propylammonium [44] (5.4–5.8 Å). The close distance between W₁₀ would contribute to the emergence of conductivity for C₁₆py-W₁₀.

C₁₆py-W₁₀ had C–H···O hydrogen bonds [40] at the interface between the W₁₀ and C₁₆py layers. The C···O distances were 3.28–3.88 Å (Table 2). The mean value was 3.51 Å, and was shorter than the mean C···O distances (~3.6 Å) in other POM hybrid crystals containing the C₁₆py cation [31,32]. These shorter hydrogen bonds as well as electrostatic interactions between C₁₆py and W₁₀ would stabilize the layered crystal structure of C₁₆py-W₁₀ with rigid packing. Most hydrogen bonds were formed between oxygen atoms of W₁₀ and the hydrophilic head of C₁₆py (i.e., pyridine rings or methylene groups near nitrogen).

Powder XRD patterns of C₁₆py-W₁₀ were measured at room temperature (Figure 3). The XRD pattern of as-prepared C₁₆py-W₁₀ exhibited weak and broad peaks (Figure 3a). The XRD pattern of recrystallized C₁₆py-W₁₀ showed much sharper and stronger peaks (Figure 3b), while the peak positions were close to those of the as-prepared C₁₆py-W₁₀ (Figure 3a). This demonstrates that the structure of as-prepared C₁₆py-W₁₀ is not changed by the recrystallization from hot acetonitrile. The pattern of the recrystallized C₁₆py-W₁₀ (Figure 3b, a = 10.7588, b = 11.5068, c = 24.7480 Å, α = 99.914, β = 93.577, γ = 116.664°, V = 2662.9 ų [45]) was almost the same as that calculated with the single crystal X-ray analysis data (Figure 3c), indicating that the recrystallized C₁₆py-W₁₀ is a single phase.

Figure 4 shows an impedance spectrum for the recrystallized C₁₆py-W₁₀ at 423 K. The spectrum showed a suppressed half circle in the high- and medium-frequency regions and an inclined line in the low-frequency region. The suppressed half circle consisted of two partially overlapped semicircles due to bulk and grain boundary resistances. The linear part in the low-frequency region would result from a combination of charge transfer resistance and Warburg impedance related to the diffusion of the carrier. The equivalent circuit [46,47,48] is shown in Figure 4: R_b and C_b are the resistance and capacitance of the bulk, respectively. R_gb and C_gb are the resistance and capacitance, respectively, of the grain boundary. R_ct and C_dl are the charge transfer resistance and double layer capacitance, respectively. Z_W is the Warburg impedance. The red line in Figure 4 represents fitted data with the equivalent circuit described above (Figure 4, inset), which successfully reproduces the measured impedance spectrum. The value of R_b obtained by the fitting was 2.25 × 10⁴ Ω, from which the conductivity of the bulk C₁₆py-W₁₀ was estimated to be 4.8 × 10⁻⁶ S cm⁻¹ considering uncertainty. The estimated value of C_b was 1.01 × 10⁻⁶ F, resulting in the time constant for the process in the bulk (R_b × C_b) of 2.27 × 10⁻² s. This short time constant suggests that the bulk process occurs by electronic conduction [47], in good agreement with the fact that C₁₆py-W₁₀ contains no easily moving ion such as H⁺. Figure 5 shows the thermogravimetric (TG) curve and IR spectra of recrystallized C₁₆py-W₁₀. No weight loss was observed below 523 K. The weight of C₁₆py-W₁₀ decreased by 36% from 523 K to 723 K (Figure 5a), which was attributed to the decomposition and removal of the C₁₆py cations. The crystallinity of C₁₆py-W₁₀ revealed by powder XRD decreased after the impedance spectroscopy measurements (not shown). However, IR spectra before and after the measurements at 423 K (Figure 5b) exhibited characteristic peaks for the W₁₀ anion in the range of 400–1000 cm⁻¹ [49], demonstrating that the molecular structure of W₁₀ was retained after heating at 423 K. These results indicate that C₁₆py-W₁₀ is thermally stable below 423 K.

The conductivity of C₁₆py-W₁₀was much lower than the radical salts of POM containing organic donor such as bis(ethylenedithio)tetrathiafulvalene [2,3]. These radical salts have conductive layers of organic donor, which possibly leads to three-dimensional conduction. On the other hand, the conductivity of C₁₆py-W₁₀is considered to be two-dimensional along the inorganic layers composed of W₁₀ and pyridinium hydrophilic heads (ab plane in the crystal). The anisotropy of the conductivity was difficult to investigate because large single crystals were not obtained. The conductivity of C₁₆py-W₁₀ reported here was measured for pelletized ground powder, and is considered to be averaged and overall conductivity. Although the conductivity of C₁₆py-W₁₀ was not so high, these results suggest that appropriate combination of POMs as electron reservoirs and surfactants with π-electrons would pave the way to another class of hybrid conductors.

All chemical reagents were obtained from commercial sources. C₁₆py-W₁₀ was synthesized according to a modified procedure of the preparation of tetrabutylammonium salt of W₁₀ [49]. 4.0 g (12.1 mmol) of Na₂WO₄·2H₂O was dissolved in 25 mL of water, and then the solution was boiled and acidified with 8.4 mL of boiling 3 M HCl solution (25 mmol) with vigorous stirring. After boiling for 2 min, 1.9 g (5.3 mmol) of [C₅H₅N(C₁₆H₃₃)]Cl·H₂O (C₁₆pyCl·H₂O) in water/ethanol (20 mL, 1:1 (v/v)) was added to form white precipitates followed by filtration and suction to dryness. Recrystallization of the crude product from hot acetonitrile gave colorless plates. The single crystals were obtained by drying up an acetonitrile solution of C₁₆py-W₁₀, which was obtained by the cation exchange of sodium salt of W₁₀ (Na-W₁₀) [34]. Data for C₁₆py-W₁₀ (single crystals): Anal. Calcd for C₈₄H₁₅₂N₄W₁₀O₃₂: C, 28.3; H, 4.3; N, 1.6%. Found: C, 28.4; H, 3.8; N, 1.6%. IR (KBr disk): 952 (m), 917 (s), 859 (m), 806 (s), 720 (w), 668 (m), 554 (w) cm⁻¹.

IR spectra (as KBr pellet) were recorded on Jasco FT-IR 5000 and Horiba FT-710 spectrometers. Thermogravimetric and differential thermal analyses (TG-DTA) were performed on an ULVAC MTS9000 + TGD9600 system. Conductivity measurements were carried out by the alternating current (AC) impedance method in a frequency range from 5 Hz to 13 MHz using an Agilent 4192A inductance-capacitance-resistance (LCR) meter. Pelletized powder samples of recrystallized C₁₆py-W₁₀ (10 mm in diameter, 0.854 mm in thickness) were sandwiched with Pt electrodes, and the impedance was measured under a dry Ar atmosphere at 423 K. Bulk resistances and conductivities of C₁₆py-W₁₀ were estimated by the fitting of typical Nyquist plots.

Single crystal X-ray diffraction measurements for C₁₆py-W₁₀ were made on a Rigaku RAXIS RAPID imaging plate diffractometer with graphite monochromated Mo-Kα radiation (λ = 0.71075 Å). Diffraction data were collected for a platelet crystal (0.30 × 0.30 × 0.02 mm) and processed with PROCESS-AUTO [50]. The structure was solved by heavy-atom Patterson methods [51] and expanded using Fourier techniques [52]. The refinement procedure was performed by the full-matrix least-squares using SHELXL97 [53]. All calculations were performed using the CrystalStructure [54] software package. Numerical absorption correction was performed for the observed data. In the refinement procedure, all non-hydrogen atoms were refined anisotropically, and the hydrogen atoms on C atoms were located in calculated positions. Further details of the crystal structure investigation may be obtained free of charge from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; Fax: +44-1223-336-033; or Email: deposit@ccdc.cam.ac.uk (CCDC-865932).

Powder X-ray diffraction (XRD) patterns for C₁₆py-W₁₀ were measured with a XRD-DSCII (Rigaku Corporation) diffractometer by using Cu Kα radiation (λ = 1.54056 Å, 50 kV-300 mA) at ambient temperature. A powder C₁₆py-W₁₀ sample was sieved in a 200 mesh sieve to remove large particles and to avoid preferred orientation. Diffraction data were collected in the range of 2θ = 2–30° at 0.01° point and 5 s/step. The lattice parameters were calculated using Materials Studio Softwares (Accelrys Inc.) by the peak profile fitting using the Pawley refinement [55].